Automate PO and Invoice

Problem: Too much manual time spent on manual work

- Atomic Projects is a start-up construction company.
- VAs needs to manually create Purchase Order in Google Sheet template, then send to Contractors
- Finance needs to manually receive Invoice from Contractors and consolidate with VAs
- The team spend 10 hour per week per each project

Approach: Turn manual job into a scaling engine

Below: Flow to upload Tally Form input into Purchase Order on Google Sheet, then upload to Notion

Below: Flow to upload Invoice sent by Contractor to Notion, using Gemini to detect Contractor and Invoice Amount

Output: Scalable machine

- Reduced 10 hours work into 1 hours work
- Allow the same VA and Finance member to scale to 10x projects
- Reduce human errors

• Back

Standardize Workflow

Problem:

- SGP Currently manages an ecosystem of 80+ Zapier flows.
- Each flow is built as a standalone, end-to-end solution for a specific scenario.
- A single schema change or API update in one system requires edit every related flow
- This creates high operational overhead, increases the risk of human error, and prevents the system from scaling efficiently.

Approach:

- Core system interactions (e.g., "Find Customer in CRM") were moved to Sub-Zaps nodes.
- Main flows now "call" these global modules, passing only the necessary project data.
- When a system changes, the assistant only edits one module, which automatically updates all related flows.

Below: Zapier flows before (left) and after (right) standarization. We reduce 70% number of active flows.

Below: An example of flow before and after clean up. We take advance of Sub-zap and Ultilities node to simplify flow.

Results:

- API and system changes are adapted within 1 hour of detection
- Enhancement cost less time to adapt

• Back

Email & Ticket AI Assistant

Problem: Important emails and tickets are not getting important attention

- The client operates a multifaceted online presence, encompassing various websites & stores
- They receive 1000+ of emails containing product and service feedback from customers.
- A substantial portion of their team's effort is consumed by the manual and time-intensive task of distinguishing genuine feedback from spam or irrelevant communications.
- Crucially, this manual filtering process often leads to critical customer communications—such as those detailing product or service disruptions, urgent negative feedback, or time-sensitive requests—not receiving the immediate attention they require.

Approach:

Step 1: Current Workflow Analysis
- The team currently utilizes Gorgias to consolidate multiple email inboxes into a single, integrated system.
- Slack serves as the primary platform for internal communication and notifications within the team.
- Previously, the team manually flagged emails in Notion based on two key criteria: their urgency (urgent/non-urgent) and their sentiment (bad, good, or neutral).
- The objective was to automate this prioritization, ensuring that emails categorized as urgent or having a negative sentiment receive immediate attention.
 
Step 2: Automated Solution with n8n & AI
- Incoming emails from Gorgias are received and processed via n8n.
- An AI agent is employed to automatically detect the urgency and sentiment of each email, and to generate a concise summary.
- Emails are automatically logged in Notion, pre-tagged with their determined urgency and sentiment.
- For emails identified as urgent or possessing a negative sentiment, immediate notifications are dispatched to the relevant Slack channel.

Results:

- The number of tickets being handled per week increased by 50%
- Critical customer feedback are response within 30 minutes

• Back

Nguyen Doan Hoang Long

[email protected] - (84) 935 209 518 - HCMC